Fix #333: allow showSelected.legend=FALSE to keep explicit showSelected without legend auto-injection

[ANAMASGARD]

Summary

Fixes #333

PR #292 added showSelected = character() so a layer can opt out of all auto legend showSelected behavior while keeping the legend visible. Issue #333 asks for a narrower case: keep an explicit showSelected (e.g. filter by year) but do not auto-add showSelected for colour/fill legend variables on the same layer (e.g. region from colour = region).

What was wrong?

On master, if you write geom_point(..., colour = region, showSelected = "year"), animint2 adds both year and region to the layer’s showSelected aesthetics. There was no way to suppress only the legend-driven part while keeping your own showSelected.

What this PR does

Commit 1 (this push): failing test only — per CONTRIBUTING

• Adds tests/testthat/test-renderer2-issue333-opt-out-ss-with-other-showSelected.R
• Uses proposed API: geom(..., showSelected = "year", showSelected.legend = FALSE)
• Asserts (once the API works):
  • Viz compiles
  • Layer showSelected is only "year" (not "region")
  • Colour legend for region is still present
  • year selector stays single for animation / tallrect
• Expected CI on commit 1: red — error at animint2dir() because error_for_showSelected_variants() does not whitelist showSelected.legend yet

Commit 2 (next): implementation

• Whitelist showSelected.legend in error_for_showSelected_variants() (R/z_animintHelpers.R)
• In addShowSelectedForLegend(), when isFALSE(extra_params$showSelected.legend), skip appending legend variables to showSelected (still honor user-specified showSelected)
• Strip showSelected.legend in getLayerParams() so it never appears in TSV/JSON export
• Expected CI after commit 2: green — same test file passes

Expected result

Users can write:

geom_point(aes(..., colour = region), showSelected = "year", showSelected.legend = FALSE)

[ANAMASGARD]

Sir @tdhock, I’ve implemented the fix for issue :

• Added showSelected.legend = FALSE to opt out of legend-driven showSelected auto-injection.
• Kept explicit showSelected = "year" working as intended.
• Preserved the existing showSelected = character() opt-out behavior from PR Fix #140: Add ability to disable automatic showSelected on legends #292.
• Whitelisted showSelected.legend in parameter validation.
• Removed showSelected.legend before export so it stays compiler-only.
• Added a regression test for the new behavior.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 73.03%. Comparing base (e447188) to head (bbb4225).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #336      +/-   ##
==========================================
+ Coverage   69.14%   73.03%   +3.89%     
==========================================
  Files         163      164       +1     
  Lines        6002     8842    +2840     
==========================================
+ Hits         4150     6458    +2308     
- Misses       1852     2384     +532     

Flag	Coverage Δ
javascript	81.25% <ø> (?)
r	69.15% <100.00%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[tdhock]

it seems reasonable to create a new argument, showSelected.legend, for this.
in that case, I would expect showSelected=character() no longer has special meaning, is that right?

[ANAMASGARD]

Yes, that is my understanding as well. I’ll consolidate on showSelected.legend=FALSE as the explicit opt-out, remove the special handling for showSelected=character(), migrate the recent tests to the new API, and note the API cleanup in NEWS.

[tdhock]

I don’t think this is necessary. Can you please revert?

[tdhock]

use informative name instead of number (plot1)